SIGNIFICANT IMPROVEMENT OF THE SHIP ENERGY EFFICIENCY VIA DESIGN OF EXPERIMENTS AND REGRESSION ANALYSIS

The main goal of the Ph.D. dissertation is to underline how a statistical approach in the planning and executive phases of the experimental activities, as well as the monitoring of complex systems can both bring real innovations in the maritime field and a higher efficiency in the results. After drawing up the Kyoto protocol, the International Maritime Organization (IMO), the UN Agency in charge for legal questions dealing with the maritime sector, has approved and ratified many measures to reduce the CO2 emissions of ships. For this reason since January 1st, 2013 a higher energy efficiency for ships is required, to be achieved through design and operation choices. During the study for the dissertation it could be shown how the application of appropriate statistical frameworks allowed to meet the goals of recent requirements. In the design step the experiment planning allowed to give designers of high speed crafts information about geometrical details of the stepped hulls. Moreover, the Design Phase has shown the strategic role that a systematic approach to planning for a design industrial experiment plays in technological process innovation. The team approach is the real driving force of pre-experimental activities. In order to predict fuel consumption and therefore carbon dioxide emission by exploiting the navigation information usually available on modern ships, a statistical model is introduced based on multiple regression analysis. For each voyage the actual fuel consumption can be compared with the consumption prediction and the prediction limits obtained through the proposed model. If the prediction interval does not included the actual fuel consumption, the management would be alerted of any change (improvement/decrease) in ship performance or the possible need for further data analysis. In fact, only with a proper and continuous monitoring of specific variables, it is possible to support sail management in making decisions. In the Operation Phase, the statistical approach presented in this thesis helps practitioners to exploit navigation information usually available on modern ships in order to predict fuel consumption, and therefore CO2 emissions, for given specific set of sailing parameters. Using these models it is possible to estimate both the reduction of fuel consumption through the improvement in energy efficiency and to estimate the CO2 emissions which is useful to get the carbon credit. During the Ph.D. a new experimental proof protocol in the towing tank test was developed: a method for the measurement of the thrust of outboard marine engines, an innovative type of construction for propeller, boat appendages and clear composite hulls to see the water flows during the experiments in the towing tank test. In this study we have shown how engineering and statistical knowledge can be integrated, how it catalyses process innovation, and, moreover how, it allows an effective cycle of step-by-step learning to be implemented in order to produce a significant improvement of the ship energy efficiency via design of experiments and regression analysis.